-
Notifications
You must be signed in to change notification settings - Fork 143
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fast softmax #972
Fast softmax #972
Conversation
…at matches its scale method (#92)
* Done to allow quantization using HQT * Added use_flash_attention and flash_attention_recompute to run_lm_eval
* Enable fast softmax mode in FusedSDPA * Add fast_softmax parameter to _gradient_checkpointing_func
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@wszczurekhabana Pls confirm we can merge this now as #967 is merged.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM!
Please run make style
.
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
Co-authored-by: Dudi Lester <160421192+dudilester@users.noreply.github.com> Co-authored-by: Sayantan Sarkar <sasarkar@habana.ai>
Changes from: HabanaAI#159
This change is dependent on: #967 to be merged first.
Original description:
Support for setting fast softmax mode in FusedSDPA operator. This is a tradeoff: performance vs accuracy.
Data on performance:
Data on accuracy (using mlperf test from: https://gerrit.habana-labs.com/plugins/gitiles/mlperf_inference/+/refs/heads/master_next/code/llama/llama_greedy.py
and https://gerrit.habana-labs.com/plugins/gitiles/mlperf_inference/+/refs/heads/master_next/code/llama/evaluation.py):